The only reservation about <…supercaps…> is their potential incompatibility with feedback controllers. These controllers depend on continually reviewing the motor speed, by measuring its back-EMF, and adjusting the width of the next few pulses in order to maintain constant motor speed. With a <…super cap…> in parallel with the motor, the perceived back-EMF will be the same as the voltage applied to the motor, more-or-less, which will affect the controller’s behaviour. It will go weird.
Would someone like to experiment and report back, please?